San Francisco Is Ground Zero for AI. How Would Congressional Candidates Push for Regulation?

TEXT START: At the center of the artificial intelligence boom, candidates looking to represent San Francisco in Washington appear aligned on a common goal: AI must be regulated.

The Dissection

This article converts systemic disruption into a campaign dispute over regulatory technique. Wiener represents controlled acceleration: transparency, liability, safety standards, unemployment expansion, and greener energy. Chan represents populist obstruction: moratoria, worker participation, liability for companies, and an end to corporate subsidies.

Both accept the same underlying premise: AI development is permanent, and the state’s task is to manage its risks. The article catalogs genuine threats—job loss, cyberattacks, privacy breaches, environmental strain, and catastrophic misuse—then funnels them into congressional procedures. Its missing subject is ownership: who controls the models, compute, energy, data, and resulting wealth.

The Core Fallacy

The text treats regulation as though it can preserve the postwar employment economy. Under the Discontinuity Thesis, it cannot.

If AI achieves durable cost and performance superiority across cognitive work, guardrails may slow deployment, assign liability, price externalities, or redistribute income. They do not restore labor’s economic necessity. A kill switch is not a wage circuit. Expanded unemployment is consumption maintenance, not productive participation. A data-center moratorium attacks infrastructure locally while competitive pressure relocates it elsewhere. Human oversight can become another layer of administratively managed labor, not a permanent human domain.

Regulation may manage the damage. It cannot reverse P1, P2, or P3.

Hidden Assumptions

• Congress can legislate faster than AI capabilities and deployment patterns change.
• U.S. regulation can constrain firms operating within global technological competition.
• Regulators will possess the technical competence and independence required to control frontier systems.
• Legal liability can be cleanly assigned when models behave unpredictably or are used by third parties.
• Human oversight will remain meaningful rather than becoming symbolic compliance theater.
• Moratoria, energy rules, and tax policy can materially alter the core automation trajectory.
• Unions and AI workers can retain bargaining power while their labor becomes increasingly replaceable.
• New AI jobs can compensate for displacement, despite the article itself citing mass layoffs and overall tech contraction.
• Public hostility toward AI can overcome the city’s dependence on the industry’s investment and tax base.
• Government regulation will be less captured than corporate self-regulation, despite the article’s own discussion of AI-funded political activity.
• Safety and employment preservation are compatible objectives rather than competing demands over who controls the transition.

Social Function

This is partial truth wrapped in transition management and ideological anesthetic. It truthfully documents political conflict and real hazards, but makes institutional responsiveness appear equivalent to structural control. Voters are offered a choice between regulated acceleration and regulated delay while ownership of AI capital remains largely untouched.

The article reassures the political class that the machine can still be placed inside a committee, a standard, a liability regime, or a bipartisan framework. That is the soothing fiction. The state may regulate the carcass, but regulation does not restore the animal.

The Verdict

This is a regulation story masquerading as a survival story. Wiener offers controlled acceleration plus redistribution; Chan offers delay, oversight, and anti-corporate anger. Both are managing the transition rather than preventing it.

Under DT logic, their proposals may create temporary lags and local niches, but neither addresses the decisive question: who owns and controls the AI capital after human labor loses its necessity? The article accurately records the political system noticing the threat. It does not recognize that the threat is not merely unsafe AI. It is the collapse of productive participation itself.